InitialAnalysisHerb
Shihong Jia
Mar 23, 2017
Load packages
library(ggplot2)
library(cowplot)
# library(lme4)
suppressWarnings(suppressMessages(library("lme4")))
library(pbkrtest)
library(tidyr)
library(dplyr)
library(optimx)
library(broom)
library(vegan)Data
herbdat <- read.csv("data/Herb_20170303.csv")
summary(herbdat)## site code quadrat
## Min. :1.000 Min. :101.0 Min. :101.0
## 1st Qu.:1.000 1st Qu.:102.0 1st Qu.:203.0
## Median :2.000 Median :104.0 Median :305.0
## Mean :2.003 Mean :103.9 Mean :300.9
## 3rd Qu.:3.000 3rd Qu.:106.0 3rd Qu.:408.0
## Max. :3.000 Max. :107.0 Max. :510.0
##
## species no.ind.
## Cardamine leucantha :1084 Min. : 1.000
## Meehania urticifolia : 817 1st Qu.: 2.000
## Brachybotrys paridiformis : 543 Median : 4.000
## Osmorhiza aristata : 538 Mean : 9.718
## Oxalis acetosella subsp. Griffithii: 440 3rd Qu.: 10.000
## Pseudostellaria sylvatica : 427 Max. :276.000
## (Other) :6547
## height coverage census treatment exclosure
## Min. : 1.000 Min. :0.0000 15fa:1295 C:2228 Min. :0.000
## 1st Qu.: 4.000 1st Qu.:0.0500 16fa:2083 F:2123 1st Qu.:0.000
## Median : 6.000 Median :0.1300 16sp:4183 I:2182 Median :1.000
## Mean : 8.881 Mean :0.2007 16su:2835 S:1634 Mean :0.504
## 3rd Qu.: 11.000 3rd Qu.:0.3200 W:2229 3rd Qu.:1.000
## Max. :126.000 Max. :0.9100 Max. :1.000
## NA's :4 NA's :132unique(herbdat$species)## [1] Filipendula palmata
## [2] Cardamine leucantha
## [3] Meehania urticifolia
## [4] Osmorhiza aristata
## [5] Athyrium brevifrons
## [6] Oxalis acetosella subsp. Griffithii
## [7] Circaea cordata
## [8] Lamium barbatum
## [9] Impatiens noli-tangere
## [10] Carex pilosa
## [11] Geum aleppicum
## [12] Arisaema amurense
## [13] Viola verecunda
## [14] Plagiorhegma dubia
## [15] Galium triflorum
## [16] Diarrhena manshurica
## [17] Brachybotrys paridiformis
## [18] Equisetum hyemale
## [19] Circaea lutetiana
## [20] Osmunda cinnamomea
## [21] Parasenecio hastatus
## [22] Paeonia obovata
## [23] Phryma leptostachya subsp. asiatica
## [24] Saussurea manshurica
## [25] Paris verticillata
## [26] Adiantum pedatum
## [27] Dryopteris crassirhizoma
## [28] Convallaria majalis
## [29] unknown spp1
## [30] Carex callitrichos
## [31] Carex siderosticta
## [32] Dioscorea nipponica
## [33] Angelica amurensis
## [34] Carex remotiuscula
## [35] unknown spp2
## [36] Adenophora remotiflora
## [37] Artemisia stolonifera
## [38] Pseudostellaria sylvatica
## [39] Maianthemum bifolium
## [40] Polemonium coeruleum
## [41] Viola acuminata
## [42] Matteuccia struthiopteris
## [43] Anemone baicalensis
## [44] Chrysosplenium sinicum
## [45] Aruncus sylvester
## [46] Doellingeria scaber
## [47] Sanicula rubriflora
## [48] Rubia cordifolia
## [49] Smilacina japonica
## [50] Actaea asiatica
## [51] Allium monanthum
## [52] Anemone spp.
## [53] Adoxa moschatellina
## [54] Corydalis repens
## [55] Anemone amurensis
## [56] Enemion raddeanum
## [57] Corydalis turtschaninovii
## [58] Lilium distichum
## [59] Anemone raddeana
## [60] Thalictrum spp.
## [61] Trillium kamtschaticum
## [62] Anemone reflexa
## [63] Ranunculus japonicus
## [64] Astilbe chinensis
## [65] Veratrum nigrum
## [66] Adonis amurensis
## [67] unknown spp4
## [68] Gagea triflora
## [69] Gagea pauciflora
## [70] Eranthis stellata
## [71] Aconitum kusnezoffii
## [72] Aconitum sczukinii
## [73] Hylomecon japonica
## [74] Corydalis yanhusuo
## [75] Urtica laetevirens
## [76] Adenocaulon himalaicum
## [77] Rabdosia excisa
## [78] unknown spp5
## [79] Chrysosplenium lectus-cochleae
## [80] Pilea pumila
## [81] Agrimonia pilosa
## [82] unknown spp3
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecundahist(table(herbdat$species))
## manipulate data
herbdat$site <- factor(herbdat$site)
herbdat$quadrat <- factor(herbdat$quadrat)
herbdat$census <- factor(herbdat$census)
herbdat$exclosure <- factor(herbdat$exclosure)
## Removing this to prevent problems later.
herbdat$species <- factor(trimws(as.character(herbdat$species)))
herbdat$quad.unique <- as.factor(paste(herbdat$site,
herbdat$exclosure,
herbdat$quadrat, sep='_'))
herbdat$plot <- as.factor(paste(herbdat$site, herbdat$exclosure,
sep='_'))
## calculate diversity
sp.diversity_herb <- summarise(
group_by(herbdat, plot, quadrat, quad.unique, treatment, exclosure, census),
shannon=diversity(no.ind.), simpson=diversity(no.ind., index='invsimpson'))
sp.diversity_herb## Source: local data frame [1,190 x 8]
## Groups: plot, quadrat, quad.unique, treatment, exclosure [?]
##
## plot quadrat quad.unique treatment exclosure census shannon
## <fctr> <fctr> <fctr> <fctr> <fctr> <fctr> <dbl>
## 1 1_0 101 1_0_101 W 0 15fa 1.1988493
## 2 1_0 101 1_0_101 W 0 16fa 1.6512596
## 3 1_0 101 1_0_101 W 0 16sp 1.3934444
## 4 1_0 101 1_0_101 W 0 16su 1.4727372
## 5 1_0 102 1_0_102 C 0 15fa 0.5623351
## 6 1_0 102 1_0_102 C 0 16fa 1.0750931
## 7 1_0 102 1_0_102 C 0 16sp 1.4268571
## 8 1_0 102 1_0_102 C 0 16su 1.3457086
## 9 1_0 103 1_0_103 I 0 15fa 1.0296530
## 10 1_0 103 1_0_103 I 0 16fa 1.2530462
## # ... with 1,180 more rows, and 1 more variables: simpson <dbl>sp.diversity_herb$census <- factor(sp.diversity_herb$census,
levels=c('15fa','16sp','16su','16fa'))
## data set of pesticide and snow removal treatment seperately
## pesticide
herbpest <- subset(herbdat, treatment !='S')
herbpest$treatment <- factor(herbpest$treatment, levels=c('W', 'F', 'I', 'C'))
sp.diversity_herbpest <- subset(sp.diversity_herb, treatment !='S')
sp.diversity_herbpest$treatment <- factor(sp.diversity_herbpest$treatment, levels=c('W', 'F', 'I', 'C'))
summary(sp.diversity_herbpest)## plot quadrat quad.unique treatment exclosure census
## 1_0:152 209 : 25 1_1_209: 5 W:237 0:472 15fa:240
## 1_1:161 504 : 25 2_1_504: 5 F:237 1:482 16sp:241
## 2_0:160 101 : 24 1_0_101: 4 I:240 16su:233
## 2_1:161 102 : 24 1_0_102: 4 C:240 16fa:240
## 3_0:160 103 : 24 1_0_103: 4
## 3_1:160 104 : 24 1_0_104: 4
## (Other):808 (Other):928
## shannon simpson
## Min. :0.000 Min. : 1.000
## 1st Qu.:1.278 1st Qu.: 2.794
## Median :1.660 Median : 3.984
## Mean :1.603 Mean : 4.353
## 3rd Qu.:1.967 3rd Qu.: 5.490
## Max. :2.973 Max. :15.979
## ## snow
herbsnow <- subset(herbdat, treatment =='C' | treatment =='S')
sp.diversity_herbsnow <- subset(sp.diversity_herb, treatment =='C' | treatment =='S')1. Analysis of diversity for pesticide
ggplot(sp.diversity_herbpest, aes(x=treatment, y=shannon, colour = exclosure)) +
geom_boxplot()
summarise(group_by(sp.diversity_herbpest, treatment, exclosure, census), meanH=mean(shannon),
seH=sd(shannon)/sqrt(length(shannon))) %>%
ggplot(aes(x=treatment, y=meanH, ymin=meanH-seH, ymax=meanH+seH, colour=census,
shape=exclosure)) + geom_pointrange() + coord_trans(y="exp") +
labs(x="pesticide", y="Effective number of species")
mod.pest.herb.div1 <- lmer(shannon ~ census*treatment*exclosure + (1|plot/quad.unique), data=sp.diversity_herbpest)
mod.pest.herb.div2 <- lmer(shannon ~ census*(treatment + exclosure) + (1|plot/quad.unique), data=sp.diversity_herbpest)
mod.pest.herb.div3 <- lmer(shannon ~ census + treatment*exclosure + (1|plot/quad.unique), data=sp.diversity_herbpest)
mod.pest.herb.div4 <- lmer(shannon ~ census*treatment + exclosure + (1|plot/quad.unique), data=sp.diversity_herbpest)
mod.pest.herb.div5 <- lmer(shannon ~ census + treatment + exclosure + (1|plot/quad.unique), data=sp.diversity_herbpest)
anova(mod.pest.herb.div1, mod.pest.herb.div2, mod.pest.herb.div3,
mod.pest.herb.div4, mod.pest.herb.div5)## refitting model(s) with ML (instead of REML)## Data: sp.diversity_herbpest
## Models:
## mod.pest.herb.div5: shannon ~ census + treatment + exclosure + (1 | plot/quad.unique)
## mod.pest.herb.div3: shannon ~ census + treatment * exclosure + (1 | plot/quad.unique)
## mod.pest.herb.div4: shannon ~ census * treatment + exclosure + (1 | plot/quad.unique)
## mod.pest.herb.div2: shannon ~ census * (treatment + exclosure) + (1 | plot/quad.unique)
## mod.pest.herb.div1: shannon ~ census * treatment * exclosure + (1 | plot/quad.unique)
## Df AIC BIC logLik deviance Chisq Chi Df
## mod.pest.herb.div5 11 854.15 907.61 -416.07 832.15
## mod.pest.herb.div3 14 859.31 927.36 -415.65 831.31 0.8379 3
## mod.pest.herb.div4 20 861.88 959.09 -410.94 821.88 9.4266 6
## mod.pest.herb.div2 23 832.69 944.49 -393.35 786.69 35.1884 3
## mod.pest.herb.div1 35 846.64 1016.77 -388.32 776.64 10.0479 12
## Pr(>Chisq)
## mod.pest.herb.div5
## mod.pest.herb.div3 0.8404
## mod.pest.herb.div4 0.1510
## mod.pest.herb.div2 1.112e-07 ***
## mod.pest.herb.div1 0.6118
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(mod.pest.herb.div2, type=c('p', 'smooth'))
dotplot(ranef(mod.pest.herb.div2, condVar=TRUE))## $`quad.unique:plot`
##
## $plot
library(sjPlot)##
## Attaching package: 'sjPlot'## The following objects are masked from 'package:cowplot':
##
## plot_grid, save_plotsjp.lmer(mod.pest.herb.div2, type='fe')## Computing p-values via Kenward-Roger approximation. Use `p.kr = FALSE` if computation takes too long.
sjp.lmer(mod.pest.herb.div2, type='re', sort.est="sort.all" )## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...
summary(mod.pest.herb.div2, correlation=FALSE)## Linear mixed model fit by REML ['lmerMod']
## Formula:
## shannon ~ census * (treatment + exclosure) + (1 | plot/quad.unique)
## Data: sp.diversity_herbpest
##
## REML criterion at convergence: 867.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.7833 -0.5039 0.0846 0.5760 2.6205
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.06172 0.2484
## plot (Intercept) 0.02127 0.1458
## Residual 0.09811 0.3132
## Number of obs: 954, groups: quad.unique:plot, 240; plot, 6
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 1.2862267 0.1020736 12.601
## census16sp 0.5759204 0.0639256 9.009
## census16su 0.6278072 0.0653146 9.612
## census16fa 0.3230473 0.0639365 5.053
## treatmentF -0.0197691 0.0728158 -0.271
## treatmentI -0.0472976 0.0729310 -0.649
## treatmentC -0.1203521 0.0728752 -1.651
## exclosure1 -0.1504535 0.1297775 -1.159
## census16sp:treatmentF 0.0490280 0.0808740 0.606
## census16su:treatmentF -0.1126312 0.0821542 -1.371
## census16fa:treatmentF -0.0464141 0.0808740 -0.574
## census16sp:treatmentI 0.0233930 0.0807355 0.290
## census16su:treatmentI -0.0231689 0.0817315 -0.283
## census16fa:treatmentI -0.0603406 0.0808740 -0.746
## census16sp:treatmentC 0.1706872 0.0808740 2.111
## census16su:treatmentC 0.0002213 0.0816560 0.003
## census16fa:treatmentC 0.0573980 0.0808740 0.710
## census16sp:exclosure1 0.3074598 0.0571376 5.381
## census16su:exclosure1 0.1070975 0.0577662 1.854
## census16fa:exclosure1 0.0291159 0.0571865 0.509##
## Correlation matrix not shown by default, as p = 20 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it2. Analysis of diversity for snow removal treatment
ggplot(sp.diversity_herbsnow, aes(x=treatment, y=shannon, colour = exclosure)) +
geom_boxplot()
summarise(group_by(sp.diversity_herbsnow, treatment, exclosure, census), meanH=mean(shannon),
seH=sd(shannon)/sqrt(length(shannon))) %>%
ggplot(aes(x=treatment, y=meanH, ymin=meanH-seH, ymax=meanH+seH, colour=census,
shape=exclosure)) + geom_pointrange() + coord_trans(y="exp") +
labs(x="pesticide", y="Effective number of species")
mod.snow.herb.div1 <- lmer(shannon ~ census*treatment*exclosure + (1|plot/quad.unique), data=sp.diversity_herbsnow)
mod.snow.herb.div2 <- lmer(shannon ~ census*(treatment + exclosure) + (1|plot/quad.unique), data=sp.diversity_herbsnow)
mod.snow.herb.div3 <- lmer(shannon ~ census + treatment*exclosure + (1|plot/quad.unique), data=sp.diversity_herbsnow)
mod.snow.herb.div4 <- lmer(shannon ~ census*treatment + exclosure + (1|plot/quad.unique), data=sp.diversity_herbsnow)
mod.snow.herb.div5 <- lmer(shannon ~ census + treatment + exclosure + (1|plot/quad.unique), data=sp.diversity_herbsnow)
anova(mod.snow.herb.div1, mod.snow.herb.div2, mod.snow.herb.div3,
mod.snow.herb.div4, mod.snow.herb.div5)## refitting model(s) with ML (instead of REML)## Data: sp.diversity_herbsnow
## Models:
## mod.snow.herb.div5: shannon ~ census + treatment + exclosure + (1 | plot/quad.unique)
## mod.snow.herb.div3: shannon ~ census + treatment * exclosure + (1 | plot/quad.unique)
## mod.snow.herb.div4: shannon ~ census * treatment + exclosure + (1 | plot/quad.unique)
## mod.snow.herb.div2: shannon ~ census * (treatment + exclosure) + (1 | plot/quad.unique)
## mod.snow.herb.div1: shannon ~ census * treatment * exclosure + (1 | plot/quad.unique)
## Df AIC BIC logLik deviance Chisq Chi Df
## mod.snow.herb.div5 9 454.31 491.80 -218.15 436.31
## mod.snow.herb.div3 10 455.68 497.34 -217.84 435.68 0.6253 1
## mod.snow.herb.div4 12 451.36 501.34 -213.68 427.36 8.3273 2
## mod.snow.herb.div2 15 442.12 504.60 -206.06 412.12 15.2329 3
## mod.snow.herb.div1 19 447.76 526.90 -204.88 409.76 2.3668 4
## Pr(>Chisq)
## mod.snow.herb.div5
## mod.snow.herb.div3 0.429083
## mod.snow.herb.div4 0.015551 *
## mod.snow.herb.div2 0.001628 **
## mod.snow.herb.div1 0.668630
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1plot(mod.snow.herb.div2, type=c('p', 'smooth'))
dotplot(ranef(mod.snow.herb.div2, condVar=TRUE))## $`quad.unique:plot`
##
## $plot
sjp.lmer(mod.snow.herb.div2, type='fe')## Computing p-values via Kenward-Roger approximation. Use `p.kr = FALSE` if computation takes too long.
sjp.lmer(mod.snow.herb.div2, type='re', sort.est="sort.all" )## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...
summary(mod.snow.herb.div2, correlation=FALSE)## Linear mixed model fit by REML ['lmerMod']
## Formula:
## shannon ~ census * (treatment + exclosure) + (1 | plot/quad.unique)
## Data: sp.diversity_herbsnow
##
## REML criterion at convergence: 456.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.05453 -0.54855 0.03804 0.62872 2.02116
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.09042 0.3007
## plot (Intercept) 0.03795 0.1948
## Residual 0.09342 0.3057
## Number of obs: 476, groups: quad.unique:plot, 122; plot, 6
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 1.18462 0.13124 9.027
## census16sp 0.78720 0.06853 11.487
## census16su 0.66114 0.06949 9.514
## census16fa 0.42863 0.06853 6.255
## treatmentS -0.29784 0.07830 -3.804
## exclosure1 -0.17043 0.17750 -0.960
## census16sp:treatmentS -0.08247 0.07957 -1.036
## census16su:treatmentS 0.13544 0.08011 1.691
## census16fa:treatmentS 0.09719 0.07957 1.221
## census16sp:exclosure1 0.22628 0.07955 2.844
## census16su:exclosure1 0.05764 0.08012 0.719
## census16fa:exclosure1 -0.06725 0.07955 -0.8453. Abundance analysis
First need to calculate abundance. In this analysis, we examine for overall, pesticide and snow removal treatment seperately, and then move to individual species.
## Overall
## data set
herb.abundat <- aggregate(no.ind. ~ treatment + exclosure + site + quadrat + plot + quad.unique + census, data=herbdat, FUN=sum)
herb.abundat$no.ind. <- as.numeric(herb.abundat$no.ind.)
mod.herb.abun1 <- glmer(no.ind. ~ site + treatment*exclosure + census +(1|quad.unique),
data=herb.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
mod.herb.abun2 <- glmer(no.ind. ~ site + treatment*exclosure + census +(1|plot/quad.unique),
data=herb.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
mod.herb.abun3 <- glmer(no.ind. ~ site + census*(treatment + exclosure) +(1|plot/quad.unique),
data=herb.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
mod.herb.abun4 <- glmer(no.ind. ~ site + census + treatment*exclosure +(1|plot/quad.unique),
data=herb.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
mod.herb.abun5 <- glmer(no.ind. ~ site + census + treatment + exclosure +(1|plot/quad.unique),
data=herb.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
anova(mod.herb.abun1, mod.herb.abun2, mod.herb.abun3,
mod.herb.abun4, mod.herb.abun5)## Data: herb.abundat
## Models:
## mod.herb.abun5: no.ind. ~ site + census + treatment + exclosure + (1 | plot/quad.unique)
## mod.herb.abun1: no.ind. ~ site + treatment * exclosure + census + (1 | quad.unique)
## mod.herb.abun2: no.ind. ~ site + treatment * exclosure + census + (1 | plot/quad.unique)
## mod.herb.abun4: no.ind. ~ site + census + treatment * exclosure + (1 | plot/quad.unique)
## mod.herb.abun3: no.ind. ~ site + census * (treatment + exclosure) + (1 | plot/quad.unique)
## Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
## mod.herb.abun5 13 15060 15126 -7516.7 15034
## mod.herb.abun1 16 15074 15156 -7521.2 15042 0.000 3 1
## mod.herb.abun2 17 15057 15144 -7511.7 15023 19.126 1 1.224e-05
## mod.herb.abun4 17 15057 15144 -7511.7 15023 0.000 0 1
## mod.herb.abun3 28 14985 15127 -7464.4 14929 94.513 11 2.164e-15
##
## mod.herb.abun5
## mod.herb.abun1
## mod.herb.abun2 ***
## mod.herb.abun4
## mod.herb.abun3 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod.herb.abun <- mod.herb.abun3
mod.herb.abun.diags <- augment(mod.herb.abun)
qplot(data=mod.herb.abun.diags, x=.fitted, y=.resid, geom=c('point', 'smooth')) +
geom_hline(yintercept=0, linetype='dotted')## `geom_smooth()` using method = 'gam'
qplot(data = mod.herb.abun.diags, x= treatment, y=.wtres, geom="boxplot", colour=census) 
sjp.lmer(mod.herb.abun, type='fe')
sjp.lmer(mod.herb.abun, type='re', sort.est="sort.all" )## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...
summary(mod.herb.abun, correlation=FALSE)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula:
## no.ind. ~ site + census * (treatment + exclosure) + (1 | plot/quad.unique)
## Data: herb.abundat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 14984.8 15127.1 -7464.4 14928.8 1162
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -6.2136 -1.5413 -0.1873 1.3159 12.0571
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.1606 0.4007
## plot (Intercept) 0.0181 0.1345
## Number of obs: 1190, groups: quad.unique:plot, 300; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.26967 0.12857 25.43 < 2e-16 ***
## site2 -0.45906 0.14629 -3.14 0.001701 **
## site3 0.02363 0.14620 0.16 0.871598
## census16fa 0.60769 0.03466 17.53 < 2e-16 ***
## census16sp 2.10014 0.02968 70.76 < 2e-16 ***
## census16su 0.99470 0.03287 30.27 < 2e-16 ***
## treatmentF 0.19784 0.06787 2.91 0.003559 **
## treatmentI -0.04490 0.07677 -0.58 0.558598
## treatmentS -0.87786 0.08040 -10.92 < 2e-16 ***
## treatmentW 0.03222 0.07795 0.41 0.679322
## exclosure1 -0.10268 0.12184 -0.84 0.399373
## census16fa:treatmentF 0.16326 0.04470 3.65 0.000260 ***
## census16sp:treatmentF 0.10428 0.03864 2.70 0.006963 **
## census16su:treatmentF 0.09675 0.04274 2.26 0.023608 *
## census16fa:treatmentI 0.16519 0.04524 3.65 0.000261 ***
## census16sp:treatmentI 0.06715 0.03917 1.71 0.086438 .
## census16su:treatmentI 0.06194 0.04312 1.44 0.150886
## census16fa:treatmentS 0.13185 0.05613 2.35 0.018816 *
## census16sp:treatmentS 0.24302 0.04844 5.02 5.26e-07 ***
## census16su:treatmentS 0.17013 0.05305 3.21 0.001341 **
## census16fa:treatmentW 0.09653 0.04497 2.15 0.031819 *
## census16sp:treatmentW 0.05238 0.03872 1.35 0.176191
## census16su:treatmentW 0.06326 0.04273 1.48 0.138758
## census16fa:exclosure1 0.08300 0.03020 2.75 0.005993 **
## census16sp:exclosure1 0.14595 0.02622 5.57 2.61e-08 ***
## census16su:exclosure1 0.13492 0.02887 4.67 2.96e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 26 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it## pesticide
herbpest.abundat <- subset(herb.abundat, treatment !='S')
herbpest.abundat$treatment <- factor(herbpest.abundat$treatment, levels=c('W', 'F', 'I', 'C'))
mod.herbpest.abun <- glmer(no.ind. ~ site + census*(treatment+exclosure)+(1|plot/quad.unique),
data=herbpest.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))## Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.mod.herbpest.abun.diags <- augment(mod.herbpest.abun)
qplot(data=mod.herbpest.abun.diags, x=.fitted, y=.resid, geom=c('point', 'smooth')) +
geom_hline(yintercept=0, linetype='dotted')## `geom_smooth()` using method = 'loess'
sjp.lmer(mod.herbpest.abun, type='fe')
sjp.lmer(mod.herbpest.abun, type='re', sort.est="sort.all" )## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.
## Plotting random effects...## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...
summary(mod.herbpest.abun, correlation=FALSE)## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula:
## no.ind. ~ site + census * (treatment + exclosure) + (1 | plot/quad.unique)
## Data: herbpest.abundat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 12402.4 12519.1 -6177.2 12354.4 930
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -6.274 -1.608 -0.227 1.419 12.320
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.128522 0.35850
## plot (Intercept) 0.009932 0.09966
## Number of obs: 954, groups: quad.unique:plot, 240; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.2976809 0.1060704 31.09 < 2e-16 ***
## site2 -0.4365160 0.1150214 -3.80 0.000148 ***
## site3 -0.0703774 0.1149445 -0.61 0.540357
## census16fa 0.7137950 0.0354978 20.11 < 2e-16 ***
## census16sp 2.1778576 0.0307451 70.84 < 2e-16 ***
## census16su 1.0736239 0.0341081 31.48 < 2e-16 ***
## treatmentF 0.1346087 0.0729559 1.85 0.065027 .
## treatmentI -0.0730403 0.0749342 -0.97 0.329696
## treatmentC -0.0523905 0.0728981 -0.72 0.472337
## exclosure1 -0.0430068 0.0972744 -0.44 0.658403
## census16fa:treatmentF 0.0663157 0.0448618 1.48 0.139348
## census16sp:treatmentF 0.0508522 0.0390387 1.30 0.192708
## census16su:treatmentF 0.0324067 0.0431566 0.75 0.452707
## census16fa:treatmentI 0.0690597 0.0453911 1.52 0.128150
## census16sp:treatmentI 0.0156481 0.0395475 0.40 0.692342
## census16su:treatmentI -0.0006187 0.0435365 -0.01 0.988661
## census16fa:treatmentC -0.0968295 0.0449688 -2.15 0.031298 *
## census16sp:treatmentC -0.0531242 0.0387193 -1.37 0.170053
## census16su:treatmentC -0.0639722 0.0427254 -1.50 0.134318
## census16fa:exclosure1 0.0622010 0.0318671 1.95 0.050952 .
## census16sp:exclosure1 0.0936266 0.0276635 3.38 0.000713 ***
## census16su:exclosure1 0.1012586 0.0305098 3.32 0.000904 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##
## Correlation matrix not shown by default, as p = 22 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it## convergence code: 0
## Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.## snow
herbsnow.abundat <- subset(herb.abundat, treatment =='C' | treatment =='S')
mod.herbsnow.abun <- glmer(no.ind. ~ site + census*treatment + exclosure+(1|plot/quad.unique),
data=herbsnow.abundat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
mod.herbsnow.abun.diags <- augment(mod.herbsnow.abun)
qplot(data=mod.herbsnow.abun.diags, x=.fitted, y=.resid, geom=c('point', 'smooth')) +
geom_hline(yintercept=0, linetype='dotted')## `geom_smooth()` using method = 'loess'
sjp.lmer(mod.herbsnow.abun, type='fe')
sjp.lmer(mod.herbsnow.abun, type='re', sort.est="sort.all" )## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.
## Plotting random effects...## Sorting each group of random intercept ('sort.all') is not possible when 'facet.grid = TRUE'.## Plotting random effects...
### individual species test
## without site as variables
## sample size data frame
sampsize.herbind <- as.data.frame(table(herbdat$species, herbdat$treatment, herbdat$exclosure))
colnames(sampsize.herbind) <- c('species', 'treatment', 'exclosure', 'no.quad')
sampsize.herbind <- sampsize.herbind %>%
spread(treatment, no.quad)
## sample size for pesticide and snow data
sampsize.herbind.pest <- sampsize.herbind[, !(colnames(sampsize.herbind) %in% c("S"))]
sampsize.herbind.snow <- sampsize.herbind[, !(colnames(sampsize.herbind) %in% c("F",'I','W'))]
sel.sp.herbpest <- unique(sampsize.herbind.pest$species[sampsize.herbind.pest$W >= 3 &
(sampsize.herbind.pest$F >= 3 | sampsize.herbind.pest$I >= 3)])
sel.sp.herbsnow <- unique(sampsize.herbind.snow$species[sampsize.herbind.snow$C >= 3 &
sampsize.herbind.snow$S >= 3])
####################
## model with census
sampsize.herbind <- as.data.frame(table(herbdat$species, herbdat$treatment, herbdat$exclosure, herbdat$census))
colnames(sampsize.herbind) <- c('species', 'treatment', 'exclosure', 'census', 'no.quad')
sampsize.herbind <- sampsize.herbind %>%
spread(treatment, no.quad)
## sample size for pesticide and snow data
sampsize.herbind.pest <- sampsize.herbind[, !(colnames(sampsize.herbind) %in% c("S"))]
sampsize.herbind.snow <- sampsize.herbind[, !(colnames(sampsize.herbind) %in% c("F",'I','W'))]
sel.sp.herbpest <- unique(sampsize.herbind.pest$species[sampsize.herbind.pest$W >= 3 &
(sampsize.herbind.pest$F >= 3 | sampsize.herbind.pest$I >= 3)])
sel.sp.herbsnow <- unique(sampsize.herbind.snow$species[sampsize.herbind.snow$C >= 3 &
sampsize.herbind.snow$S >= 3])
indivmods.herbpest <- sapply(sel.sp.herbpest, function(sp){
print(sp)
spdat <- filter(herbpest, species == sp)
spdat <- droplevels(spdat)
# random effects: remove quad.unique
if (length(unique(spdat$exclosure)) == 2 & length(unique(spdat$census)) >= 2){
mod <- glmer(no.ind. ~ census + treatment + exclosure + (1|plot/quad.unique),
data=spdat,family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
} else if (length(unique(spdat$exclosure)) < 2 & length(unique(spdat$census)) >=2) {
mod <- glmer(no.ind. ~ census + treatment + (1|plot/quad.unique),
data=spdat,family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
} else if (length(unique(spdat$exclosure)) == 2 & length(unique(spdat$census)) < 2) {
mod <- glmer(no.ind. ~ treatment + exclosure + (1|plot/quad.unique),
data=spdat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
} else {
mod <- glmer(no.ind. ~ treatment + (1|plot/quad.unique),
data=spdat, family=poisson,
control=glmerControl(optimizer="optimx",
optCtrl=list(method=c('bobyqa', 'Nelder-Mead'))))
}
}, simplify=FALSE)## [1] Adenophora remotiflora
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Adonis amurensis
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Adoxa moschatellina
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control
## $checkConv, : Model failed to converge with max|grad| = 0.0508559 (tol =
## 0.001, component 1)## [1] Allium monanthum
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Anemone amurensis
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Anemone baicalensis
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Anemone raddeana
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Anemone spp.
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Arisaema amurense
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Athyrium brevifrons
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Brachybotrys paridiformis
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Cardamine leucantha
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.## [1] Carex pilosa
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Carex remotiuscula
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.## [1] Circaea cordata
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Circaea lutetiana
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Corydalis repens
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Corydalis turtschaninovii
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Diarrhena manshurica
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Enemion raddeanum
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Equisetum hyemale
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Eranthis stellata
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Filipendula palmata
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Gagea pauciflora
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Gagea triflora
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Galium triflorum
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Impatiens noli-tangere
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Lamium barbatum
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Lilium distichum
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.## [1] Maianthemum bifolium
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Meehania urticifolia
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Osmorhiza aristata
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Oxalis acetosella subsp. Griffithii
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Parasenecio hastatus
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.## [1] Paris verticillata
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control
## $checkConv, : Model failed to converge with max|grad| = 0.0436602 (tol =
## 0.001, component 1)## [1] Phryma leptostachya subsp. asiatica
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda## Warning in optimx.check(par, optcfg$ufn, optcfg$ugr, optcfg$uhess, lower, : Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.## [1] Pseudostellaria sylvatica
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Ranunculus japonicus
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Saussurea manshurica
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Smilacina japonica
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Thalictrum spp.
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] unknown spp2
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] unknown spp4
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda
## [1] Viola verecunda
## 82 Levels: Aconitum kusnezoffii Aconitum sczukinii ... Viola verecunda### plot residuals
lapply(indivmods.herbpest, function(mod){
plot(mod, type=c('p', 'smooth'))})## [[1]]
##
## [[2]]
##
## [[3]]
##
## [[4]]
##
## [[5]]
##
## [[6]]
##
## [[7]]
##
## [[8]]
##
## [[9]]
##
## [[10]]
##
## [[11]]
##
## [[12]]
##
## [[13]]
##
## [[14]]
##
## [[15]]
##
## [[16]]
##
## [[17]]
##
## [[18]]
##
## [[19]]
##
## [[20]]
##
## [[21]]
##
## [[22]]
##
## [[23]]
##
## [[24]]
##
## [[25]]
##
## [[26]]
##
## [[27]]
##
## [[28]]
##
## [[29]]
##
## [[30]]
##
## [[31]]
##
## [[32]]
##
## [[33]]
##
## [[34]]
##
## [[35]]
##
## [[36]]
##
## [[37]]
##
## [[38]]
##
## [[39]]
##
## [[40]]
##
## [[41]]
##
## [[42]]
##
## [[43]]
##
## [[44]]
lapply(indivmods.herbpest, function(mod){
dotplot(ranef(mod, condVar=TRUE))})## [[1]]
## [[1]]$`quad.unique:plot`
##
## [[1]]$plot
##
##
## [[2]]
## [[2]]$`quad.unique:plot`
##
## [[2]]$plot
##
##
## [[3]]
## [[3]]$`quad.unique:plot`
##
## [[3]]$plot
##
##
## [[4]]
## [[4]]$`quad.unique:plot`
##
## [[4]]$plot
##
##
## [[5]]
## [[5]]$`quad.unique:plot`
##
## [[5]]$plot
##
##
## [[6]]
## [[6]]$`quad.unique:plot`
##
## [[6]]$plot
##
##
## [[7]]
## [[7]]$`quad.unique:plot`
##
## [[7]]$plot
##
##
## [[8]]
## [[8]]$`quad.unique:plot`
##
## [[8]]$plot
##
##
## [[9]]
## [[9]]$`quad.unique:plot`
##
## [[9]]$plot
##
##
## [[10]]
## [[10]]$`quad.unique:plot`
##
## [[10]]$plot
##
##
## [[11]]
## [[11]]$`quad.unique:plot`
##
## [[11]]$plot
##
##
## [[12]]
## [[12]]$`quad.unique:plot`
##
## [[12]]$plot
##
##
## [[13]]
## [[13]]$`quad.unique:plot`
##
## [[13]]$plot
##
##
## [[14]]
## [[14]]$`quad.unique:plot`
##
## [[14]]$plot
##
##
## [[15]]
## [[15]]$`quad.unique:plot`
##
## [[15]]$plot
##
##
## [[16]]
## [[16]]$`quad.unique:plot`
##
## [[16]]$plot
##
##
## [[17]]
## [[17]]$`quad.unique:plot`
##
## [[17]]$plot
##
##
## [[18]]
## [[18]]$`quad.unique:plot`
##
## [[18]]$plot
##
##
## [[19]]
## [[19]]$`quad.unique:plot`
##
## [[19]]$plot
##
##
## [[20]]
## [[20]]$`quad.unique:plot`
##
## [[20]]$plot
##
##
## [[21]]
## [[21]]$`quad.unique:plot`
##
## [[21]]$plot
##
##
## [[22]]
## [[22]]$`quad.unique:plot`
##
## [[22]]$plot
##
##
## [[23]]
## [[23]]$`quad.unique:plot`
##
## [[23]]$plot
##
##
## [[24]]
## [[24]]$`quad.unique:plot`
##
## [[24]]$plot
##
##
## [[25]]
## [[25]]$`quad.unique:plot`
##
## [[25]]$plot
##
##
## [[26]]
## [[26]]$`quad.unique:plot`
##
## [[26]]$plot
##
##
## [[27]]
## [[27]]$`quad.unique:plot`
##
## [[27]]$plot
##
##
## [[28]]
## [[28]]$`quad.unique:plot`
##
## [[28]]$plot
##
##
## [[29]]
## [[29]]$`quad.unique:plot`
##
## [[29]]$plot
##
##
## [[30]]
## [[30]]$`quad.unique:plot`
##
## [[30]]$plot
##
##
## [[31]]
## [[31]]$`quad.unique:plot`
##
## [[31]]$plot
##
##
## [[32]]
## [[32]]$`quad.unique:plot`
##
## [[32]]$plot
##
##
## [[33]]
## [[33]]$`quad.unique:plot`
##
## [[33]]$plot
##
##
## [[34]]
## [[34]]$`quad.unique:plot`
##
## [[34]]$plot
##
##
## [[35]]
## [[35]]$`quad.unique:plot`
##
## [[35]]$plot
##
##
## [[36]]
## [[36]]$`quad.unique:plot`
##
## [[36]]$plot
##
##
## [[37]]
## [[37]]$`quad.unique:plot`
##
## [[37]]$plot
##
##
## [[38]]
## [[38]]$`quad.unique:plot`
##
## [[38]]$plot
##
##
## [[39]]
## [[39]]$`quad.unique:plot`
##
## [[39]]$plot
##
##
## [[40]]
## [[40]]$`quad.unique:plot`
##
## [[40]]$plot
##
##
## [[41]]
## [[41]]$`quad.unique:plot`
##
## [[41]]$plot
##
##
## [[42]]
## [[42]]$`quad.unique:plot`
##
## [[42]]$plot
##
##
## [[43]]
## [[43]]$`quad.unique:plot`
##
## [[43]]$plot
##
##
## [[44]]
## [[44]]$`quad.unique:plot`
##
## [[44]]$plot
lapply(indivmods.herbpest, function(x) summary(x, correlation=FALSE))## [[1]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 396.3 418.5 -188.2 376.3 58
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.3896 -0.7003 -0.3194 0.5528 2.1206
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.562 0.7497
## plot (Intercept) 0.000 0.0000
## Number of obs: 68, groups: quad.unique:plot, 34; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.1932 0.7948 -0.243 0.807894
## census16fa 1.4054 0.7264 1.935 0.053011 .
## census16sp 2.2337 0.7194 3.105 0.001901 **
## census16su 2.6442 0.7180 3.683 0.000231 ***
## treatmentF 0.3359 0.3828 0.878 0.380196
## treatmentI -0.2622 0.4426 -0.592 0.553666
## treatmentC 0.2975 0.4219 0.705 0.480689
## exclosure1 -1.1350 0.3280 -3.461 0.000539 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[2]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 231.2 243.9 -108.6 217.2 38
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.24302 -0.68936 0.03475 0.51888 1.15953
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.2019 0.4493
## plot (Intercept) 0.0000 0.0000
## Number of obs: 45, groups: quad.unique:plot, 45; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.06652 0.24667 4.324 1.53e-05 ***
## treatmentF 0.59801 0.29169 2.050 0.0403 *
## treatmentI 0.54540 0.30607 1.782 0.0748 .
## treatmentC 0.39864 0.28265 1.410 0.1584
## exclosure1 -0.03868 0.19983 -0.194 0.8465
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[3]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1453.7 1487.6 -717.8 1435.7 311
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.61687 -0.53196 -0.08948 0.37939 2.68756
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.246061 0.49605
## plot (Intercept) 0.002631 0.05129
## Number of obs: 320, groups: quad.unique:plot, 163; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.051998 0.001418 36.7 <2e-16 ***
## census16sp 1.214466 0.001418 856.5 <2e-16 ***
## census16su 1.000188 0.001418 705.5 <2e-16 ***
## treatmentF 0.251437 0.001417 177.4 <2e-16 ***
## treatmentI 0.318203 0.001417 224.5 <2e-16 ***
## treatmentC 0.154555 0.001417 109.0 <2e-16 ***
## exclosure1 0.029006 0.001418 20.5 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Model failed to converge with max|grad| = 0.0508559 (tol = 0.001, component 1)
##
##
## [[4]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1391.8 1414.2 -688.9 1377.8 175
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.26042 -0.33517 0.00582 0.18330 0.33528
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.62654 0.7915
## plot (Intercept) 0.06826 0.2613
## Number of obs: 182, groups: quad.unique:plot, 182; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.46699 0.21105 11.689 <2e-16 ***
## treatmentF 0.28454 0.18147 1.568 0.117
## treatmentI -0.06145 0.18214 -0.337 0.736
## treatmentC -0.01336 0.17877 -0.075 0.940
## exclosure1 -0.07854 0.24936 -0.315 0.753
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[5]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 2435.7 2459.9 -1210.9 2421.7 226
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.75531 -0.14240 0.01566 0.08460 0.20769
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.7202 0.8487
## plot (Intercept) 0.1230 0.3506
## Number of obs: 233, groups: quad.unique:plot, 233; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 4.0827 0.2394 17.053 <2e-16 ***
## treatmentF -0.1200 0.1610 -0.745 0.456
## treatmentI -0.1570 0.1609 -0.976 0.329
## treatmentC -0.0646 0.1600 -0.404 0.686
## exclosure1 -0.2479 0.3080 -0.805 0.421
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[6]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 791.9 817.0 -385.9 771.9 81
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.8030 -1.1121 -0.3899 0.6558 5.1412
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.7416 0.8612
## plot (Intercept) 0.0000 0.0000
## Number of obs: 91, groups: quad.unique:plot, 34; plot, 3
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.4328 0.4139 1.046 0.296
## census16fa 0.9866 0.2158 4.571 4.85e-06 ***
## census16sp 2.8923 0.1957 14.779 < 2e-16 ***
## census16su 3.2737 0.1949 16.798 < 2e-16 ***
## treatmentF 0.2980 0.4436 0.672 0.502
## treatmentI -0.3831 0.4168 -0.919 0.358
## treatmentC -0.3491 0.4157 -0.840 0.401
## exclosure1 -0.1249 0.3181 -0.393 0.694
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[7]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 590.3 609.8 -287.2 574.3 76
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.24459 -0.28497 -0.02441 0.22054 0.45976
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.5669 0.7529
## plot (Intercept) 0.0000 0.0000
## Number of obs: 84, groups: quad.unique:plot, 83; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.64643 1.07580 0.601 0.548
## census16sp 1.49300 1.05422 1.416 0.157
## treatmentF 0.13113 0.27969 0.469 0.639
## treatmentI 0.30633 0.25198 1.216 0.224
## treatmentC -0.04721 0.25262 -0.187 0.852
## exclosure1 -0.02352 0.19811 -0.119 0.905
##
## [[8]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 170.3 179.1 -78.1 156.3 19
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.18169 -0.40673 0.09073 0.28263 0.64728
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.4823 0.6945
## plot (Intercept) 0.0000 0.0000
## Number of obs: 26, groups: quad.unique:plot, 26; plot, 2
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.61705 0.32227 5.018 5.23e-07 ***
## treatmentF 0.50351 0.42561 1.183 0.237
## treatmentI 0.50689 0.42585 1.190 0.234
## treatmentC -0.43236 0.48357 -0.894 0.371
## exclosure1 0.04706 0.33168 0.142 0.887
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[9]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 297.7 318.8 -138.9 277.7 51
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.47053 -0.51244 -0.09982 0.41978 2.07621
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.21929 0.4683
## plot (Intercept) 0.09572 0.3094
## Number of obs: 61, groups: quad.unique:plot, 29; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.87853 0.35001 5.367 8.00e-08 ***
## census16fa -0.97916 0.22884 -4.279 1.88e-05 ***
## census16sp -0.52181 0.16400 -3.182 0.00146 **
## census16su 0.22807 0.13591 1.678 0.09333 .
## treatmentF -0.86181 0.30199 -2.854 0.00432 **
## treatmentI -0.83382 0.45314 -1.840 0.06575 .
## treatmentC -1.14623 0.38258 -2.996 0.00274 **
## exclosure1 0.08685 0.36775 0.236 0.81331
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[10]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 255.3 281.5 -117.7 235.3 91
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.53991 -0.25004 -0.19831 0.02704 1.43182
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0 0
## plot (Intercept) 0 0
## Number of obs: 101, groups: quad.unique:plot, 50; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.15353 0.24826 0.618 0.536
## census16fa -0.14567 0.26787 -0.544 0.587
## census16sp 0.09562 0.25385 0.377 0.706
## census16su 0.09586 0.24574 0.390 0.696
## treatmentF 0.01306 0.30137 0.043 0.965
## treatmentI 0.20481 0.25113 0.816 0.415
## treatmentC -0.03490 0.26131 -0.134 0.894
## exclosure1 0.07936 0.20317 0.391 0.696
##
## [[11]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 2306.3 2347.3 -1143.1 2286.3 438
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1596 -0.5486 -0.1378 0.4045 3.3871
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.6707 0.8190
## plot (Intercept) 0.1958 0.4425
## Number of obs: 448, groups: quad.unique:plot, 157; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.56591 0.30435 1.859 0.063 .
## census16fa 0.74335 0.07709 9.642 <2e-16 ***
## census16sp 1.28434 0.07287 17.626 <2e-16 ***
## census16su 1.27029 0.07409 17.145 <2e-16 ***
## treatmentF -0.10269 0.20494 -0.501 0.616
## treatmentI 0.08965 0.20174 0.444 0.657
## treatmentC -0.08390 0.19493 -0.430 0.667
## exclosure1 -0.42943 0.39247 -1.094 0.274
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[12]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 5819.7 5867.8 -2899.8 5799.7 897
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4861 -0.8254 -0.1137 0.6283 8.0141
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.2081 0.4561
## plot (Intercept) 0.1155 0.3398
## Number of obs: 907, groups: quad.unique:plot, 237; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.516592 0.209980 7.223 5.1e-13 ***
## census16fa 0.652098 0.029934 21.785 < 2e-16 ***
## census16sp 0.706036 0.029585 23.865 < 2e-16 ***
## census16su 0.529590 0.030727 17.235 < 2e-16 ***
## treatmentF 0.186653 0.088761 2.103 0.0355 *
## treatmentI 0.175275 0.089792 1.952 0.0509 .
## treatmentC -0.000753 0.089653 -0.008 0.9933
## exclosure1 0.355648 0.284645 1.249 0.2115
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
##
##
## [[13]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1688.4 1726.6 -834.2 1668.4 327
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5181 -0.5681 -0.2027 0.3797 3.8899
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.6483 0.8052
## plot (Intercept) 0.0745 0.2729
## Number of obs: 337, groups: quad.unique:plot, 99; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.06121 0.24995 4.246 2.18e-05 ***
## census16fa 0.25452 0.06763 3.763 0.000168 ***
## census16sp 0.39441 0.06630 5.949 2.70e-09 ***
## census16su 0.48459 0.06543 7.406 1.30e-13 ***
## treatmentF 0.16925 0.24509 0.691 0.489839
## treatmentI -0.04732 0.24481 -0.193 0.846716
## treatmentC -0.31457 0.25351 -1.241 0.214667
## exclosure1 0.02077 0.28664 0.072 0.942243
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[14]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 457.8 484.0 -218.9 437.8 92
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.1758 -0.4776 -0.2514 0.3632 3.5140
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 5.671e-01 7.531e-01
## plot (Intercept) 6.461e-16 2.542e-08
## Number of obs: 102, groups: quad.unique:plot, 47; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.0006436 0.3817867 -0.002 0.998655
## census16fa 0.4229696 0.1761388 2.401 0.016335 *
## census16sp 0.6425333 0.1664055 3.861 0.000113 ***
## census16su 0.5212396 0.1760831 2.960 0.003074 **
## treatmentF 0.7001961 0.3739002 1.873 0.061112 .
## treatmentI 0.4659222 0.4066670 1.146 0.251915
## treatmentC 0.2652807 0.4591589 0.578 0.563431
## exclosure1 -0.1844836 0.2759277 -0.669 0.503755
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
##
##
## [[15]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 516.8 540.4 -249.4 498.8 93
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4146 -0.4890 -0.2041 0.3228 3.3416
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.5469 0.7395
## plot (Intercept) 0.0000 0.0000
## Number of obs: 102, groups: quad.unique:plot, 85; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.24188 1.08625 -0.223 0.8238
## census16fa 1.26387 1.08452 1.165 0.2439
## census16su 1.54344 1.07608 1.434 0.1515
## treatmentF 0.11730 0.27892 0.420 0.6741
## treatmentI -0.04357 0.28990 -0.150 0.8805
## treatmentC -0.19707 0.28332 -0.696 0.4867
## exclosure1 -0.37475 0.20785 -1.803 0.0714 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[16]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 448.0 474.3 -214.0 428.0 92
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5822 -0.5450 -0.1479 0.3632 1.8044
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.3729 0.6107
## plot (Intercept) 0.0000 0.0000
## Number of obs: 102, groups: quad.unique:plot, 57; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.8319 0.2383 3.492 0.00048 ***
## census16fa -0.4416 0.2422 -1.823 0.06829 .
## census16sp -1.1130 0.2607 -4.270 1.95e-05 ***
## census16su 0.2254 0.1224 1.841 0.06564 .
## treatmentF 0.4427 0.3145 1.408 0.15923
## treatmentI 0.2382 0.2774 0.859 0.39055
## treatmentC 0.1388 0.3099 0.448 0.65424
## exclosure1 -0.1934 0.2128 -0.909 0.36346
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[17]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1871.1 1894.7 -928.5 1857.1 209
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.82777 -0.23149 0.02501 0.15341 0.41889
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.4487 0.6699
## plot (Intercept) 0.2413 0.4912
## Number of obs: 216, groups: quad.unique:plot, 216; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.9102 0.3058 9.518 <2e-16 ***
## treatmentF 0.2054 0.1387 1.481 0.139
## treatmentI 0.1852 0.1374 1.347 0.178
## treatmentC 0.2019 0.1380 1.463 0.144
## exclosure1 0.1219 0.4129 0.295 0.768
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[18]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1147.8 1173.9 -565.9 1131.8 185
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.35606 -0.44858 -0.09649 0.27186 1.01838
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.4113 0.6413
## plot (Intercept) 0.1701 0.4125
## Number of obs: 193, groups: quad.unique:plot, 189; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.496709 0.438669 1.132 0.25750
## census16sp 1.005545 0.339429 2.962 0.00305 **
## treatmentF 0.193078 0.157676 1.224 0.22075
## treatmentI 0.004239 0.160718 0.026 0.97896
## treatmentC 0.316229 0.159704 1.980 0.04769 *
## exclosure1 0.120181 0.361494 0.332 0.73954
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[19]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 420.1 446.8 -200.0 400.1 97
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.2358 -0.5267 -0.1009 0.4238 2.1813
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.09759 0.3124
## plot (Intercept) 0.07865 0.2804
## Number of obs: 107, groups: quad.unique:plot, 38; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.94602 0.29890 3.165 0.001551 **
## census16fa 0.43601 0.16589 2.628 0.008582 **
## census16sp 0.34518 0.16475 2.095 0.036156 *
## census16su 0.57385 0.15736 3.647 0.000266 ***
## treatmentF -0.78230 0.26096 -2.998 0.002719 **
## treatmentI 0.10917 0.21307 0.512 0.608388
## treatmentC -0.58940 0.25678 -2.295 0.021714 *
## exclosure1 -0.09217 0.30322 -0.304 0.761151
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[20]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 398.0 420.1 -190.0 380.0 77
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3837 -0.5763 -0.2547 0.3128 1.8627
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.4198 0.6479
## plot (Intercept) 0.0000 0.0000
## Number of obs: 86, groups: quad.unique:plot, 47; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.8803 0.3536 2.490 0.01279 *
## census16sp 0.7994 0.2541 3.145 0.00166 **
## census16su 0.7488 0.2519 2.973 0.00295 **
## treatmentF -0.7385 0.3593 -2.056 0.03982 *
## treatmentI -0.7533 0.3169 -2.377 0.01745 *
## treatmentC -0.7411 0.3341 -2.218 0.02656 *
## exclosure1 0.0936 0.2492 0.376 0.70716
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[21]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 812.3 847.7 -396.2 792.3 243
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.3643 -0.4603 -0.1474 0.3077 4.4097
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.1 0.3163
## plot (Intercept) 0.0 0.0000
## Number of obs: 253, groups: quad.unique:plot, 106; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.406547 0.168164 2.418 0.01563 *
## census16fa 0.426732 0.145646 2.930 0.00339 **
## census16sp 0.045427 0.168414 0.270 0.78736
## census16su 0.347470 0.147283 2.359 0.01831 *
## treatmentF -0.165827 0.175827 -0.943 0.34562
## treatmentI -0.006441 0.151502 -0.042 0.96609
## treatmentC 0.158244 0.155209 1.020 0.30794
## exclosure1 -0.239882 0.121035 -1.982 0.04749 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[22]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1092.9 1118.0 -538.4 1076.9 163
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.36956 -0.44994 -0.02384 0.25681 2.23279
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.58779 0.7667
## plot (Intercept) 0.02272 0.1507
## Number of obs: 171, groups: quad.unique:plot, 158; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.4316 0.1958 12.421 < 2e-16 ***
## census16su -1.0351 0.1590 -6.509 7.56e-11 ***
## treatmentF -0.4905 0.1995 -2.458 0.0140 *
## treatmentI -0.3097 0.1954 -1.585 0.1130
## treatmentC -0.1698 0.1984 -0.856 0.3921
## exclosure1 -0.3483 0.1947 -1.789 0.0737 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[23]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1116.2 1149.4 -548.1 1096.2 195
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1002 -0.7797 -0.2281 0.5473 5.9680
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.35126 0.5927
## plot (Intercept) 0.04874 0.2208
## Number of obs: 205, groups: quad.unique:plot, 63; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.91113 0.25093 3.631 0.000282 ***
## census16fa 0.25810 0.09705 2.660 0.007825 **
## census16sp 0.91129 0.08503 10.717 < 2e-16 ***
## census16su 0.19852 0.09572 2.074 0.038075 *
## treatmentF 0.37161 0.23466 1.584 0.113282
## treatmentI 0.26893 0.23007 1.169 0.242440
## treatmentC 0.23311 0.24399 0.955 0.339372
## exclosure1 -0.26008 0.28984 -0.897 0.369550
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[24]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 677.3 694.4 -331.6 663.3 79
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.83772 -0.31638 -0.00503 0.13073 0.28653
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.9296 0.9642
## plot (Intercept) 0.1771 0.4209
## Number of obs: 86, groups: quad.unique:plot, 86; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.47709 0.40236 6.156 7.44e-10 ***
## treatmentF 0.12513 0.32110 0.390 0.697
## treatmentI 0.00742 0.32446 0.023 0.982
## treatmentC -0.09743 0.31091 -0.313 0.754
## exclosure1 -0.48860 0.45873 -1.065 0.287
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[25]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 917.8 935.9 -451.9 903.8 90
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.02578 -0.22425 0.02181 0.08840 0.22665
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.9977 0.9989
## plot (Intercept) 0.2590 0.5089
## Number of obs: 97, groups: quad.unique:plot, 97; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 3.37986 0.43852 7.707 1.28e-14 ***
## treatmentF -0.03696 0.30356 -0.122 0.903
## treatmentI 0.01052 0.30494 0.034 0.972
## treatmentC -0.13928 0.29507 -0.472 0.637
## exclosure1 -0.63895 0.51862 -1.232 0.218
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[26]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1306.1 1341.1 -643.1 1286.1 235
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5698 -0.6979 -0.2716 0.5267 3.6836
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.5044 0.7102
## plot (Intercept) 0.0000 0.0000
## Number of obs: 245, groups: quad.unique:plot, 73; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.86255 0.21419 4.027 5.65e-05 ***
## census16fa 0.57119 0.08653 6.601 4.08e-11 ***
## census16sp 1.01998 0.08025 12.709 < 2e-16 ***
## census16su 0.66204 0.08635 7.667 1.76e-14 ***
## treatmentF 0.25668 0.25102 1.023 0.307
## treatmentI 0.30859 0.25626 1.204 0.229
## treatmentC 0.14443 0.25786 0.560 0.575
## exclosure1 -0.24376 0.18144 -1.343 0.179
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[27]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1549.2 1586.3 -764.6 1529.2 292
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.1426 -0.6623 -0.2713 0.5493 4.1255
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.2614 0.5112
## plot (Intercept) 0.1027 0.3205
## Number of obs: 302, groups: quad.unique:plot, 115; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.73492 0.31080 2.365 0.018 *
## census16fa 0.21424 0.15315 1.399 0.162
## census16sp 0.81354 0.14315 5.683 1.32e-08 ***
## census16su 0.91411 0.14198 6.438 1.21e-10 ***
## treatmentF -0.06830 0.15528 -0.440 0.660
## treatmentI -0.21628 0.15811 -1.368 0.171
## treatmentC -0.21779 0.15106 -1.442 0.149
## exclosure1 0.06945 0.34084 0.204 0.839
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[28]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 252.2 271.5 -116.1 232.2 41
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.09730 -0.55654 -0.08424 0.32248 1.64176
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.3401 0.5831
## plot (Intercept) 0.0000 0.0000
## Number of obs: 51, groups: quad.unique:plot, 35; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.72850 0.40342 4.285 1.83e-05 ***
## census16fa -0.46234 0.34695 -1.333 0.1827
## census16sp 0.36255 0.60775 0.597 0.5508
## census16su 0.01212 0.29199 0.042 0.9669
## treatmentF -0.37011 0.35388 -1.046 0.2956
## treatmentI -0.06169 0.45361 -0.136 0.8918
## treatmentC 0.07508 0.34541 0.217 0.8279
## exclosure1 -0.54254 0.26530 -2.045 0.0409 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[29]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 466.6 490.0 -224.3 448.6 90
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.6922 -0.5343 -0.3512 0.3043 2.4551
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.5122 0.7157
## plot (Intercept) 0.0000 0.0000
## Number of obs: 99, groups: quad.unique:plot, 70; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.32745 0.41462 -0.790 0.430
## census16sp 1.21991 0.30538 3.995 6.48e-05 ***
## census16su 1.26755 0.30545 4.150 3.33e-05 ***
## treatmentF -0.18670 0.32422 -0.576 0.565
## treatmentI -0.23109 0.33265 -0.695 0.487
## treatmentC -0.07917 0.31578 -0.251 0.802
## exclosure1 0.12679 0.22729 0.558 0.577
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
##
##
## [[30]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 943.6 976.8 -461.8 923.6 193
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.5080 -0.5334 -0.1544 0.2482 2.3298
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.6794 0.8243
## plot (Intercept) 0.0000 0.0000
## Number of obs: 203, groups: quad.unique:plot, 93; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.19556 0.74334 -1.608 0.107758
## census16fa 1.78384 0.71691 2.488 0.012838 *
## census16sp 2.45628 0.71479 3.436 0.000590 ***
## census16su 2.37541 0.71492 3.323 0.000892 ***
## treatmentF -0.35364 0.27597 -1.281 0.200040
## treatmentI -0.19453 0.27471 -0.708 0.478870
## treatmentC 0.07528 0.25437 0.296 0.767278
## exclosure1 0.06576 0.19683 0.334 0.738322
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[31]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 5094.6 5140.8 -2537.3 5074.6 736
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8240 -0.7520 -0.1286 0.6838 5.6426
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.8039 0.8966
## plot (Intercept) 0.1097 0.3312
## Number of obs: 746, groups: quad.unique:plot, 218; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.59569 0.24102 6.621 3.57e-11 ***
## census16fa 0.67614 0.03431 19.710 < 2e-16 ***
## census16sp 1.01412 0.03288 30.846 < 2e-16 ***
## census16su 0.76758 0.03441 22.306 < 2e-16 ***
## treatmentF -0.15405 0.17390 -0.886 0.376
## treatmentI -0.25253 0.17854 -1.414 0.157
## treatmentC -0.07753 0.16991 -0.456 0.648
## exclosure1 -0.02563 0.29868 -0.086 0.932
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[32]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 2622.7 2664.5 -1301.3 2602.7 473
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5130 -0.7168 -0.1679 0.4744 4.5669
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.45287 0.673
## plot (Intercept) 0.02857 0.169
## Number of obs: 483, groups: quad.unique:plot, 173; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.67036 0.16838 3.981 6.86e-05 ***
## census16fa 1.01894 0.07771 13.112 < 2e-16 ***
## census16sp 0.80273 0.08001 10.033 < 2e-16 ***
## census16su 1.17841 0.07665 15.375 < 2e-16 ***
## treatmentF 0.03023 0.15825 0.191 0.848
## treatmentI -0.12507 0.16025 -0.780 0.435
## treatmentC -0.04981 0.15411 -0.323 0.747
## exclosure1 -0.08124 0.18054 -0.450 0.653
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[33]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 2576.2 2615.5 -1278.1 2556.2 364
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -5.1081 -0.8340 -0.2325 0.5387 7.6897
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.7443 0.8627
## plot (Intercept) 0.1065 0.3264
## Number of obs: 374, groups: quad.unique:plot, 121; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.80501 0.25732 7.015 2.3e-12 ***
## census16fa 0.52834 0.04380 12.063 < 2e-16 ***
## census16sp -0.59608 0.05910 -10.087 < 2e-16 ***
## census16su 0.01491 0.04904 0.304 0.7611
## treatmentF -0.52044 0.24540 -2.121 0.0339 *
## treatmentI -0.25108 0.23535 -1.067 0.2861
## treatmentC -0.00968 0.22498 -0.043 0.9657
## exclosure1 -0.21234 0.34632 -0.613 0.5398
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[34]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 246.9 267.5 -113.5 226.9 48
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1712 -0.4479 -0.1177 0.2456 1.8952
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.2385 0.4883
## plot (Intercept) 0.0000 0.0000
## Number of obs: 58, groups: quad.unique:plot, 28; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.97986 0.58991 -1.661 0.096709 .
## census16fa 1.42909 0.53144 2.689 0.007165 **
## census16sp 1.58312 0.52823 2.997 0.002726 **
## census16su 1.91969 0.52285 3.672 0.000241 ***
## treatmentF 0.75901 0.37978 1.999 0.045657 *
## treatmentI 0.05891 0.39017 0.151 0.879984
## treatmentC 0.21187 0.33084 0.640 0.521909
## exclosure1 0.18100 0.27437 0.660 0.509450
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
##
##
## [[35]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 769.6 799.6 -374.8 749.6 138
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.87801 -0.59860 -0.02626 0.31673 1.88947
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 4.557e-01 0.6750686
## plot (Intercept) 2.033e-08 0.0001426
## Number of obs: 148, groups: quad.unique:plot, 106; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.039266 0.001637 24.0 <2e-16 ***
## census16fa -0.037781 0.001635 -23.1 <2e-16 ***
## census16sp 1.606745 0.001637 981.5 <2e-16 ***
## census16su 0.290372 0.001635 177.6 <2e-16 ***
## treatmentF -0.014153 0.001636 -8.7 <2e-16 ***
## treatmentI 0.183950 0.001636 112.4 <2e-16 ***
## treatmentC 0.038881 0.001636 23.8 <2e-16 ***
## exclosure1 0.003387 0.001637 2.1 0.0385 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Model failed to converge with max|grad| = 0.0436602 (tol = 0.001, component 1)
##
##
## [[36]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 215.8 233.2 -98.9 197.8 42
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.9748 -0.5488 -0.2258 0.2114 1.5869
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.1921 0.4383
## plot (Intercept) 0.0000 0.0000
## Number of obs: 51, groups: quad.unique:plot, 41; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.72428 0.49952 1.450 0.1471
## census16fa 0.58484 0.42723 1.369 0.1710
## census16su 0.34953 0.43729 0.799 0.4241
## treatmentF 0.10343 0.40117 0.258 0.7965
## treatmentI -0.06969 0.33273 -0.210 0.8341
## treatmentC -0.03335 0.34747 -0.096 0.9235
## exclosure1 -0.55898 0.23687 -2.360 0.0183 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## convergence code: 0
## Parameters or bounds appear to have different scalings.
## This can cause poor performance in optimization.
## It is important for derivative free methods like BOBYQA, UOBYQA, NEWUOA.
##
##
## [[37]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 1443.7 1482.0 -711.8 1423.7 330
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.9036 -0.5516 -0.1287 0.3094 3.3025
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.2945 0.5426
## plot (Intercept) 0.1098 0.3314
## Number of obs: 340, groups: quad.unique:plot, 170; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.20104 1.07408 0.187 0.852
## census16fa 0.17542 1.06211 0.165 0.869
## census16sp 0.85556 1.05832 0.808 0.419
## census16su 0.97918 1.05906 0.925 0.355
## treatmentF -0.02294 0.15363 -0.149 0.881
## treatmentI -0.18000 0.14700 -1.224 0.221
## treatmentC -0.04480 0.14935 -0.300 0.764
## exclosure1 0.03831 0.29237 0.131 0.896
##
## [[38]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 513.1 534.3 -248.6 497.1 96
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1356 -0.4775 -0.1764 0.3991 1.1174
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.28393 0.5329
## plot (Intercept) 0.04642 0.2155
## Number of obs: 104, groups: quad.unique:plot, 103; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.49457 0.50856 0.972 0.331
## census16sp 0.67473 0.46207 1.460 0.144
## treatmentF 0.06742 0.22538 0.299 0.765
## treatmentI 0.14440 0.20139 0.717 0.473
## treatmentC 0.14406 0.20546 0.701 0.483
## exclosure1 0.12906 0.23017 0.561 0.575
##
## [[39]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 210.3 231.1 -95.2 190.3 49
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.8695 -0.4751 -0.1201 0.2784 2.0151
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.1296 0.3601
## plot (Intercept) 0.0000 0.0000
## Number of obs: 59, groups: quad.unique:plot, 26; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.2481969 0.6873264 -0.361 0.7180
## census16fa 0.4961348 0.6463335 0.768 0.4427
## census16sp 1.1263778 0.6302712 1.787 0.0739 .
## census16su 1.0602241 0.6425455 1.650 0.0989 .
## treatmentF -0.0005728 0.3284512 -0.002 0.9986
## treatmentI -0.4627426 0.4615937 -1.002 0.3161
## treatmentC 0.2566551 0.3293613 0.779 0.4358
## exclosure1 -0.0647791 0.2408050 -0.269 0.7879
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[40]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 284.2 304.1 -132.1 264.2 44
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.91485 -0.46320 -0.04718 0.39584 1.29404
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 1.02560 1.0127
## plot (Intercept) 0.01845 0.1358
## Number of obs: 54, groups: quad.unique:plot, 27; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.3287 1.1320 -1.174 0.24046
## census16fa 1.5245 1.0486 1.454 0.14601
## census16sp 3.2876 1.0183 3.229 0.00124 **
## census16su 2.9149 1.0295 2.831 0.00464 **
## treatmentF 0.1851 0.6727 0.275 0.78326
## treatmentI -0.9168 1.2068 -0.760 0.44744
## treatmentC -0.4440 0.7843 -0.566 0.57133
## exclosure1 -0.6273 0.5622 -1.116 0.26456
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[41]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 326.8 347.8 -154.4 308.8 67
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.0176 -0.4514 -0.1487 0.2348 1.6057
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.40685 0.6379
## plot (Intercept) 0.06955 0.2637
## Number of obs: 76, groups: quad.unique:plot, 38; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.6372 0.3908 1.631 0.1030
## census16sp 0.4353 0.1953 2.229 0.0258 *
## census16su 0.4842 0.1955 2.477 0.0132 *
## treatmentF 0.1058 0.3941 0.269 0.7883
## treatmentI 0.3676 0.3821 0.962 0.3359
## treatmentC -0.3545 0.3757 -0.944 0.3453
## exclosure1 -0.1642 0.3819 -0.430 0.6672
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[42]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 302.1 326.2 -141.1 282.1 72
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -0.93573 -0.46663 -0.07868 0.33451 1.78963
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.2062 0.4541
## plot (Intercept) 0.0000 0.0000
## Number of obs: 82, groups: quad.unique:plot, 46; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.03789 0.49821 0.076 0.9394
## census16fa 0.36163 0.43452 0.832 0.4053
## census16sp 0.30640 0.43271 0.708 0.4789
## census16su 0.53355 0.42953 1.242 0.2142
## treatmentF 0.33052 0.34681 0.953 0.3406
## treatmentI 0.48659 0.30043 1.620 0.1053
## treatmentC 0.53673 0.31669 1.695 0.0901 .
## exclosure1 -0.29822 0.22657 -1.316 0.1881
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[43]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 161.1 175.0 -72.5 145.1 34
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.2145 -0.6554 -0.1027 0.6683 2.0972
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.05459 0.2336
## plot (Intercept) 0.00000 0.0000
## Number of obs: 42, groups: quad.unique:plot, 40; plot, 5
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.74580 0.32590 2.288 0.0221 *
## census16su -0.53419 0.36870 -1.449 0.1474
## treatmentF 0.01234 0.40387 0.030 0.9756
## treatmentI 0.52210 0.33381 1.564 0.1178
## treatmentC 0.05081 0.35508 0.143 0.8862
## exclosure1 -0.10303 0.22975 -0.448 0.6538
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## [[44]]
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: poisson ( log )
## Formula: no.ind. ~ census + treatment + exclosure + (1 | plot/quad.unique)
## Data: spdat
## Control:
## glmerControl(optimizer = "optimx", optCtrl = list(method = c("bobyqa",
## "Nelder-Mead")))
##
## AIC BIC logLik deviance df.resid
## 497.2 527.6 -238.6 477.2 145
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1256 -0.5191 -0.2870 0.2742 2.2798
##
## Random effects:
## Groups Name Variance Std.Dev.
## quad.unique:plot (Intercept) 0.027415 0.16558
## plot (Intercept) 0.005977 0.07731
## Number of obs: 155, groups: quad.unique:plot, 78; plot, 6
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.52256 0.18194 2.872 0.00408 **
## census16fa -0.05325 0.17892 -0.298 0.76600
## census16sp -0.32291 0.26322 -1.227 0.21991
## census16su 0.44374 0.15560 2.852 0.00435 **
## treatmentF 0.01956 0.17475 0.112 0.91089
## treatmentI -0.21008 0.19612 -1.071 0.28407
## treatmentC 0.02537 0.18650 0.136 0.89179
## exclosure1 -0.21845 0.17335 -1.260 0.20760
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Overall, pesticide treatments didn’t affect diversity but abundance. However, when density was tested in individual species, pesticide caused some effects. Snow removal treatment changed both abundance and diversity. I haven’t done the analyses of snow removal on the density of individual species. As to the analysis of pestcide treatment on density of individual species, results were complicated but interesting.